Impact of determination method and sample type on basic wood density analysis and its implications for Eucalyptus pulp production

2.1 Sampling of material

The sampled trees were obtained from an experimental clonal test plot located in São Miguel Arcanjo, São Paulo, Brazil, at coordinates 23° 52′ 7.7″ S and 47° 53′ 45.5″ W. According to the Köppen classification, the region has a Cfa climate, characterized as humid subtropical (Alvares et al. 2013), and the soil is classified as yellow latosol. The planting area covers 3.56 ha, with a spacing of 3.3 × 1.8 m, and all plots underwent similar silvicultural management. Between 2016 and 2022, the region had an average annual temperature of 20.2 °C and precipitation of 1,100.86 mm.

The clonal test was conducted using a completely randomized block design, with linear plots of five plants distributed across six blocks, totaling 480 plots. Five operational Eucalyptus clones relevant to the partner company were collected at 6.8 years (6 years and 10 months) of age. Based on the data from the last forest inventory conducted in the experimental plots, five trees with an average diameter were selected from each clone, totaling 25 sampled trees, whose dendrometric information is presented in Table 1.

After harvesting, a rigorous cubing of the trees was carried out using the Smalian method (Machado and Figueiredo Filho 2009), considering the commercial height defined for the region, where the diameter with bark was limited to 6 cm. Subsequently, disks 3 cm thick were cut at six positions along the trunk of each tree: 0, DAP (1.30 m), 25 %, 50 %, 75 %, and 100 % of the commercial trunk height (Figure 1A). In addition, five bolts of 1 m in length were collected between the sampling positions of the disks to produce wood chips (Figure 1C).

Figure 1:

Sampling of discs and wood chips for basic density determination. MMC: maximum moisture content.

From the disks collected at six positions from the base to the top, wedges were produced (Figure 1B) for the determination of basic density. To produce wood chips, the five bolts cut from each tree were chipped as individual samples and classified using a shaker sieve with a mesh opening of 25 mm, resulting in approximately 85 % classified chips. The final classification was carried out manually to remove chips with non-standard dimensions, knots, and cracks.

After classification, samples of 60 g of chips were collected from each tree, and their dimensions were measured with a steel caliper. The average dimensions of the chips were approximately 25 mm in width, 25 mm in length, and 3 mm in thickness. Subsequently, the chips were subjected to natural drying for 48 h to prevent the colonization of staining and blocking fungi during transportation and storage until the tests were conducted.

2.2 Determination of basic density by wedges

From each disk along the trunk, a wedge was removed (Figure 1B) for the determination of basic wood density, according to NBR 11941 (ABNT 2003), using vacuum to accelerate the saturation of the samples. The basic density of each tree was determined based on the average of the wedges collected at the six positions.

The criterion adopted to consider the samples saturated was constant mass, which consists of performing consecutive weighing of the samples. The material is considered saturated when the variation in mass between weighing is less than 0.5 g.

2.3 Determination of basic density of wood chips and saturation tests

The basic density of wood chips was determined using the maximum moisture content method, according to ABNT NBR 11941 (2003) and the hydrostatic balance method, employing different times and saturation procedures (Table 2). The hydrostatic balance method is described in the SCAN-CM 43:45 (1995), ABNT NBR 11941 (2003), TAPPI 258 OM-02 (2006), and ASTM D2395-14 (2017) standards. Three samples of 130 g of dry chips (per repetition) were collected from each tree, with an initial average moisture content of 9 %, totaling 75 chip samples per test. All tests were conducted in a vertical autoclave (Primatec CS-A-150L), with a maximum working capacity of 127 °C and a pressure of 0.18 MPa.

The verification of the saturation of wood chip samples was conducted according to two criteria: i) control of constant wet mass, which involves weighing the sample on a precision digital scale until the mass becomes constant between consecutive weighing, with the difference between them not exceeding 0.5 g (ABNT NBR 11941 2003; SCAN-CM 43:45 1995); and ii) submersion of the chips in water. The samples were immersed in a container with water, observing whether they floated or submerged. The chips were considered saturated when they were submerged. Both criteria were adopted because they are recognized methodologies used in laboratories and companies in the forestry sector.

After verifying the condition of the chips and confirming that they were saturated according to both criteria, the samples were dried with cotton towels and absorbent paper to remove excess water from their surface before weighing the wet mass and determining the volume.

The variables of temperature, vacuum, and pressure were employed to reduce the saturation time of the chips for density determination, and they are widely adopted by various companies in the forestry sector. For this purpose, three samples of chips (triplicate) from each genetic material were maintained in a closed system (autoclave) under the influence of the variables listed in Table 2.

During the tests with the variables and their combinations, the mass of the chip samples was measured every three days until a constant mass was reached, defined as the situation where the difference between consecutive weighing does not exceed 0.5 g. For Tests 2 and 3, following the protocol of the partner company, weighing commenced after four days, continuing at intervals of 4, 6, and 9 days, and so on, until a constant mass was obtained. The volume of the samples was measured from the moment the samples were completely submerged in water.

After the chips reached a constant mass, this value was recorded for the determination of basic density using the maximum moisture content method. Subsequently, the samples were immersed in water using a stainless-steel basket (Figure 1C) for the determination of basic density by the hydrostatic balance method.

During weighing the chip samples every three days, it was possible to observe the behavior of moisture and basic density, as well as to evaluate the saturation time. It was also possible to assess the effect of saturation tests and genetic material on these parameters. The limit of 10 kg/m³ was chosen to compare the differences in the results of the basic density of the wood, as this is the value considered impactful by companies in the pulp sector.

2.4 Specific wood consumption

To assess the importance of basic wood density values in the pulp mill context, the specific wood consumption was estimated based on the different basic density results obtained considering submersion and complete saturation of samples. This calculation also incorporated pulping yield and losses values reported in the literature, derived from cooking processes with kappa numbers ranging from 15.5 to 16.7, conducted using the same hybrids and with average basic density values similar to those obtained in the present study (Table 3). This estimation aims to evaluate the impact of variations in basic density – resulting from different measurement methods – on specific wood consumption.

Equation (1) presents the formula used to estimate laboratory-scale specific wood consumption (SWC). The K factor (900) represents the conversion to one air-dried metric ton of pulp, assuming 10 % moisture content. Basic density refers to the laboratory-measured wood density, while GY (Global Yield) corresponds to the pulp yield after deducting process losses, which can range from 6 % to 9 %. This study does not consider actual biological variations in basic density but rather the differences arising from the use of different measurement methods. Nevertheless, when assuming a fixed specific wood consumption, variations in basic density may reflect changes in the actual global yield, which is challenging to measure accurately, especially due to uncertainties associated with process losses.

where: GY = (PY × (1 – losses))

where: SWC: specific wood consumption (m³.adt⁻¹); Factor K: 900, conversion factor for adt (ton of air-dried pulp); BD: basic wood density (kg/m³); PY: pulping yield (%).

2.5 Data analysis

To analyze the basic density data among the different days and tests, a statistical approach was adopted that included the comparison of means and standard deviations. The experimental design was carried out in a completely randomized model, ensuring that the observed variability was attributed solely to the applied treatments. Initially, an analysis of variance (ANOVA) was performed to verify significant differences among the sample means, using the F test to identify relevant variations. When ANOVA indicated statistical significance, a multiple comparison Tukey test was conducted, with a significance level of 5 %, allowing for the identification of which groups exhibited specific differences among themselves and providing a more detailed analysis of the observed variations.

3.1 Moisture and basic density of wood (maximum moisture content method) over the days of water saturation

The results regarding the moisture and basic density of the wood chips, evaluated using the maximum moisture content method over the weighing days, are presented in Figures 2 and 3. A decrease in the basic density of the wood was observed over the saturation days, while the moisture content increased. Furthermore, the genetic material had an impact on the saturation time of the samples by the submersion method, with variations in saturation days among the clones in tests (1, 4, and 5). However, when measuring saturation by constant mass (with variations less than 0.5 g), it was found that the clones saturated on the same day within the same test.

Figure 2:

Variation in moisture, starting at 9 %, and basic density (maximum moisture content) of wood chips over the evaluation days.

Figure 3:

Variation in moisture, starting at 9 %, and basic density (maximum moisture content) of wood chips over the evaluation days.

Table 4 presents the values of basic density and moisture content of the wood chips measured on the day of saturation, using both the water submersion criterion (common in companies and laboratories) and the constant wet control, determined by consecutive weighing. The cells highlighted in blue indicate differences greater than 10 kg/m³.

Basic density varied over the days of saturation using the maximum moisture content method (Figures 2 and 3), highlighting the importance of ensuring complete saturation of the samples. This control was based on the stabilization of the mass between successive weighing. It was observed that the submersion control method, widely used in industries, overestimated the basic density, as the samples continued to gain mass even after complete submersion, which impacted the density results.

Clone C3 showed the highest average basic density, followed by clones C2, C1, C5, and C4, respectively. The order was reversed for moisture levels, suggesting an inverse correlation between density and moisture, consistently observed across all tests conducted.

The basic density of the wood chips was evaluated both on the day of complete submersion and on the final saturation day when the mass variation between weighing was less than 0.5 g. Significant differences in density values were identified in tests 1, 2, 4, 5, and 6 (Table 4) for at least one clone. The magnitude of these differences varied according to the genetic material and the test, with clone C5 showing the largest discrepancy between the saturation criteria, reaching a difference of 28 kg/m³ in test 4. However, this trend was not observed in tests 3, 6, and 7 for the same clone.

The tests that combined vacuum, pressure, and temperature (Tests 2 and 3), considered more severe due to the association of these variables, showed a shorter saturation time for the wood chips and smaller differences between basic density values when comparing saturated and submerged samples. In Test 3, which was the most severe (with higher vacuum, pressure, and temperature), the differences ranged from 0 kg/m³ for clone C4 to 4 kg/m³ for clone C5, without these differences being statistically significant.

On the other hand, Tests 5 and 6, which were less severe due to not including temperature in the saturation process, required more time to reach wood saturation (6 and 9 days for submersion saturation and 15 days for constant mass saturation). These tests also showed the largest differences between the saturation criteria, with variations exceeding 10 kg/m³, and these differences were statistically significant for most of the clones analyzed.

Among the seven tests performed, Tests 2 and 3 resulted in a saturation time of 6 days, as determined by consecutive weighing until reaching constant mass. Test 7 had a saturation time of 9 days, while Tests 1 and 4 took 12 days. In Tests 5 and 6, constant mass saturation occurred after 15 days.

To determine the basic density of wood using the hydrostatic balance method, where the chips must be completely submerged in water for volume measurement (submersion criterion), these times were reduced. In Tests 2 and 3, the saturation time decreased to 4 days, while in Tests 1, 4, 5, 6, and 7, submersion occurred between 6 and 9 days, varying according to the genetic material.

3.2 Comparison of basic wood density determined with wedges and chips: hydrostatic balance method

In the comparative analysis between the basic density values of wedges and chips, determined by the hydrostatic balance method (Figure 4), the wedges showed higher average basic density values compared to the chips. The differences varied from 0 kg/m³ for clone C3 to 29 kg/m³ for clone C2, with this difference being statistically significant only for clone C2.

Figure 4:

Basic wood density (hydrostatic balance method) determined with wedges and chips. Bars followed by the same letter within the same clone do not show statistical differences among themselves. Clones without letters presented in the graphs also do not differ statistically from each other, according to the Tukey test at 5 % significance (P<0.05). *: Greater differences between the basic densities determined by wedge and chip. **: Smaller differences between the basic densities determined by wedge and chip.

Clone C3, which exhibited the lowest wood density among the genetic materials studied, also had the smallest variations between the density values of wedges and chips, with a maximum variation of 6 kg/m³. These results indicate that the difference between the two sampling methods is less pronounced in materials with lower basic density, such as clone C3, while materials with higher basic density, like clone C2, may show more pronounced variations between samples.

The standard deviation of the basic density, when determined with wedges, was higher than that of the chips, indicating greater variability in the results obtained from the wedges. However, this trend was not repeated for clones C2 and C1, whose standard deviations were lower in the wedge samples (12 kg/m³ and 9 kg/m³, respectively) compared to the chips in test 1, which showed deviations of 19 kg/m³ for C2 and 14 kg/m³ for C1.

When analyzing the overall averages of the clones, the trend observed individually was maintained: the basic density determined with wedges showed higher absolute values compared to the chips. The differences between the samples, although present in absolute values ranging from 1 to 11 kg/m³, were not statistically significant, as indicated by the Tukey test at 5 % significance. This suggests that, despite the variations, the precision and reproducibility of the methods remain similar.

3.3 Basic density of chips: comparison of hydrostatic and maximum moisture content methods

The basic density of the chips, when determined by the two methods – hydrostatic balance and maximum moisture content – with saturation using only daily vacuum and saturation criterion based on constant mass, did not show significant differences (Figure 5). In absolute terms, the variation between the methods was minimal, with clone C5 showing a difference of 2 kg/m³ and clone C2 showing a difference of 7 kg/m³. The maximum moisture content method resulted in slightly higher values of basic density compared to the hydrostatic balance, reinforcing the consistency between the methods, although small variations may occur depending on the genetic material.

Figure 5:

Comparison of the basic density of chips obtained by the hydrostatic balance and maximum moisture content methods. The means followed by the same letter in the column, within the same clone, do not differ statistically from each other according to Tukey’s test at 5 % significance (P<0.05). MMC: maximum moisture content.

In terms of the overall average among the clones, the difference between the hydrostatic balance and maximum moisture content methods also showed no significant difference, as observed in the individually analyzed clones. The overall average difference among the clones was 5 kg/m³, indicating that, despite small variations in the basic density values between the methods, these variations were not statistically relevant.

3.4 Estimation of specific wood consumption by different saturation methods for chips and sample types

Specific wood consumption is a crucial factor for the pulp mill, as it reflects the amount of wood needed to produce one ton of pulp. It is one of the main indicators of raw material usage efficiency and, therefore, directly influences production costs and demand for wood, significantly impacting the sector. In this study, variations in the values of basic density and yield of different genetic materials were analyzed. Based on these variables, it was possible to estimate the impact of these variations on specific wood consumption, as shown in Figure 6.

Figure 6:

Comparison of estimated specific wood consumption with different basic densities (maximum moisture content). Numbers in the bars correspond to specific wood consumption (m³/ton of pulp).

The relationship between basic density and specific wood consumption (SWC), often mentioned in the literature, was confirmed in this study. Clones with higher basic density, such as C1 and C4, exhibited a lower SWC, meaning they required less wood to produce one ton of pulp. This behavior reflects the efficiency of genetic materials with higher density in the pulp production process, as they tend to offer greater yield per unit volume of wood. On the other hand, clones with lower basic density, such as C2 and C3, showed a higher SWC, indicating the need for a larger volume of wood to achieve the same result in terms of pulp production.

The differences observed in SWC resulting from the different methods of saturation assessment (samples saturated to constant mass versus submerged samples) did not follow a specific pattern for each clone. However, if these differences were reflected in the industrial reality, the impact would be significant.

Additionally, the saturation method influenced the measurement of basic density, which may have been underestimated, leading to an overestimation of global yield in the predictions of specific wood consumption over the evaluation days. The tests that required more time for wood saturation resulted in greater variations in basic density, which in turn caused overestimates in global yield, without directly impacting specific wood consumption, as observed in tests 1, 4, 5, and 6. The opposite was true for tests 2, 3, and 7, which exhibited smaller variations in basic density and, consequently, less impact on global yield predictions.

4.1 Humidity and basic density of wood (maximum moisture content method) over saturation days and test effects

The analysis of the results indicates an inverse relationship between basic density and wood moisture. Clones that exhibited higher basic density also showed lower moisture percentages, corroborating observations made in previous studies on Eucalyptus wood (Almeida et al. 2014; Foelkel et al. 1983; Oliveira et al. 2005; Silva 2019).

This relationship is observed in both the natural moisture content of the wood and its maximum moisture content. The inverse correlation occurs due to the reduced volume of void spaces within the wood’s anatomical structure, which limits its capacity for water storage. In other words, changes in the morphology of fibers and vessels in trees lead to different distributions of vessels and fibers that vary in length and wall thickness. This structural arrangement results in an increase in basic density and a decrease in cellular spaces, where moisture – mostly in the form of free water – tends to accumulate (Carmo 1996; Oliveira et al. 2005; Silva 2019; Zobel and Van Buijtenen 1989).

The effects of chip saturation tests on density values highlight the importance of ensuring complete saturation of samples through consecutive weighing, as well as emphasizing the risks associated with measuring saturation using the bucket immersion methodology. As mentioned by Foelkel et al. (1971), it is crucial to carefully control the water absorption of samples and ensure that maximum saturation is achieved for the determination of density using the maximum moisture content method.

The determination of basic wood density by the maximum moisture content method is based on two fundamental assumptions: first, the wood must be at its maximum moisture content during the weighing of saturated samples; second, the density of the “wood substance” must be 1,530 kg/m³ for the first condition to be met (Foelkel et al. 1983). This methodology has been widely adopted in the wood sector due to its operational efficiency, as it involves only weighing wet and dry chips in an oven, followed by calculations based on a specific equation. In contrast, the hydrostatic balance method requires more equipment, and the process of measuring the volume of samples is more time-consuming, which can be a limiting factor in industrial environments that seek to optimize production.

Studies, including the present work, highlight the risk of overestimating the basic wood density when measuring it without ensuring maximum saturation. In comparing the maximum moisture content method with the hydrostatic balance method, Foelkel et al. (1971) emphasizes the practicality of ensuring only the saturation point of the fibers and the lower likelihood of measurement errors associated with the hydrostatic balance method.

The decrease observed in the curve of basic density values over the days until the complete saturation of the material can be explained by the formula used to determine basic density by the maximum moisture content method. In this calculation, the presence of the wet mass in the numerator of the fraction that forms the denominator of the formula results in an inversely proportional relationship between the wet mass and the basic density of the wood. As the wet mass increases during the saturation process, the calculated basic density tends to decrease, reflecting the importance of ensuring complete saturation to obtain an accurate estimate of basic wood density.

No significant effect of the saturation test on the final basic density values was observed. The small variations found in absolute terms were expected due to the high variability of wood, which can occur both within the same tree and among trees of the same stand, age, species, and genus (Ferreira and Kageyama 1978; Foelkel et al. 1992).

The time required for wood saturation to determine basic density is often a limiting factor for companies in the forestry sector. An extended saturation period for samples can delay analyses and processes, as well as complicate “in situ” monitoring of the raw material used in production and the generation of information for forestry improvement programs. Therefore, the use of different tests for wood saturation aimed to accelerate this process and analyze the effect of this acceleration on the final basic density values of the wood.

It was anticipated that the combination of variables in the conducted tests would result in a shorter saturation time for the samples. However, this expectation was not confirmed for all tests. Tests 5 (vacuum and temperature) and 6 (pressure and temperature) had a longer final saturation time than tests 1 (vacuum) and 4 (pressure).

4.2 Basic density of wedges and chips

The basic density of wood in the samples of wedges and chips showed significant differences in average values. Previous studies, such as those by Segura (2015) and Coelho (2017; 2021), which determined the basic density of disks and chips from Corymbia, Pinus, and Eucalyptus, corroborate these findings. These authors also observed that larger samples, such as disks, exhibited higher absolute values of basic density compared to chips. For example, Segura (2015) recorded variations from 48 to 1,140 kg/m³ for disks and chips of Corymbia and Eucalyptus. Coelho (2017) reported that the basic density of chips from Pinus was 4.5–5.4 % lower than that of disks, while Coelho (2021) obtained average results that were 4.1 % lower than the basic density of Eucalyptus chips.

These significant variations in basic wood density, resulting from differences in samples, highlight the importance of studies that deepen the understanding of basic density determination. In addition to influencing the physical characterization of wood, these variations can directly impact the prediction of specific consumption, reflecting on the efficiency and economy of production processes in the timber and pulp industries.

4.3 Method for determining the basic density of wood: maximum moisture content and hydrostatic balance

The basic density of chips, determined by the maximum moisture content method, was consistently higher for all clones compared to the hydrostatic balance method. However, no statistically significant differences were found between the two methods, as analyzed by the Tukey test at a 5 % significance level.

Previous studies, such as those by Foelkel et al. (1971), which compared methods for determining basic density in chips from conifer and broadleaf species, also did not identify significant differences. Similarly, Padula (2013) observed comparable results when investigating the two methods in Eucalyptus wood.

Although the maximum moisture content method is more practical and feasible for large-scale determinations, it is crucial to maintain strict control over the moisture of the samples, ensuring proper saturation. In this regard, the hydrostatic balance method stands out for being less influenced by variables that can impact the precision of measurements, as highlighted by Foelkel et al. (1971).

4.4 Estimation of specific consumption by different saturation methods for wood chips and sample types

The predictions of specific wood consumption were significantly influenced by the various parameters studied, including the saturation tests and criteria, the types of samples, and the method of density determination. These different measures directly impacted on the estimation of this index, highlighting the importance of obtaining accurate basic density values that reflect the reality of the factory.

By ensuring accurate measurements, it is possible to avoid underestimating or overestimating specific wood consumption, allowing for a better understanding of loss points along the production chain. Foelkel (2015) emphasizes the relevance of paying attention to the calculations of specific wood consumption, as variations in values may conceal important information about disturbances in the process and about the intrinsic and extrinsic quality of the wood.

Although the range of variation in the basic density of wood among the clones studied in this work was not high, a significant trend was observed in estimating specific wood consumption. Using materials with higher density resulted in lower wood demand, while the opposite occurred with clones of lower density. This trend is supported by Mokfienski et al. (2008) and Gomide et al. (2010), who emphasize that wood is often marketed and transported by volume. Therefore, denser wood tends to contain more cellulose in a smaller volume, leading to reduced wood demand. This advantage, in terms of raw material utilization efficiency, can be considered in forestry improvement programs.

Although the relationship between low basic wood density and specific consumption may seem unfavorable, it is important to recognize the advantages associated with lower-density woods. For example, these woods require a lower alkaline load during cooking, which can result in better yield, viscosity, and lower solid content for recovery. This aspect makes it essential to study the optimal point between basic density, specific consumption, and the production process (Gomide et al. 2005).

The trends observed in the clones regarding specific consumption remained consistent across different tests, indicating that, in addition to basic density, genetic material and refined yield have a significant influence on the predictions of specific wood consumption. This finding underscores the importance of selecting the appropriate raw material and implementing forestry improvement programs that consider these variables.

The comparison between C2 and C3 clones revealed an expected behavior in the relationship between basic density and specific wood consumption. Clone C3, with lower basic density, showed higher predictive values of specific consumption compared to clone C2. A study conducted by Gomide et al. (2010) with 75 Eucalyptus clones aimed to correlate the basic characteristics of wood with the yield of pulping. It was found that, although there is a high correlation between the basic density of wood and specific consumption, factors such as lignin and extractives content also have a significant impact on the yield. These components can cause variations that affect the specific wood consumption, resulting in losses or gains that need to be considered in the analysis.

Guerra et al. (2008) identified a strong correlation between yield and specific wood consumption in clones of Eucalyptus globulus aged between 5 and 7 years, all cultivated in the same region. This geographical uniformity suggests that growing conditions play a significant role in the wood characteristics. On the other hand, Mokfienski et al. (2008) found that while lower basic density is associated with higher yield, the specific wood consumption is lower in Eucalyptus species with higher density. This observation suggests that even though less dense wood may yield more, it requires a larger amount of wood for cellulose production, contradicting the idea that higher yield always results in lower specific consumption.

Another study conducted by Magaton et al. (2009) concluded that in their analysis, specific wood consumption is more influenced by basic density than by yield. Additionally, they observed that yield was not associated with other wood properties, indicating that density may be a critical factor to consider. Sansígolo and Ramos (2011) corroborated the idea that higher basic density is associated with higher yield and lower specific wood consumption. This relationship reinforces the need for careful analysis when selecting clones for forest improvement programs, considering not only density but also yield and other factors that may impact specific wood consumption.